type-directed, whitespace-delimited parsing for embedded dsls

Type-Directed, Whitespace-De-limited Parsing for Embedded DSLs

Cyrus Omar

School of Computer Science Carnegie Mellon University

[GlobalDSL13]

Benjamin Chung Alex PotaninDarya Kurilova Jonathan Aldrich

http://cs.c-mu.edu/~dkurilov/

Carnegie Mellon University, School of Computer Science

2

Wyvern Goals: Secure web and mobile programming within a

single statically-typed language.

Language-level support for a variety of domains: Security policies and architecture specifications Client-side programming (HTML, CSS) Server-side programming (Databases)

Carnegie Mellon University, School of Computer Science

3

Benefits of DSLs Specialized syntax improves ease-of-use Specialized typechecking rules improve verifiability Specialized translation strategies improve performance

and interoperability with existing technologies Specialized tool support improves ease-of-use.

Carnegie Mellon University, School of Computer Science

4

Types of DSLs Standalone DSLs are external; must call into each other

via interoperability layers.

Embedded DSLs use mechanisms internal to a host general-purpose language; distributed and accessed as libraries.

Carnegie Mellon University, School of Computer Science

5

Natural Interoperability

Carnegie Mellon University, School of Computer Science

6

Caveat: Expressivity vs. Safety Want expressive (syntax) extensions. But if you give each DSL too much control, they may

interfere with one another at link-time.

Carnegie Mellon University, School of Computer Science

7

Example: SugarJ Libraries can extend the base syntax of the language These extensions are imported transitively Even seemingly simple extensions can interfere:

Pairs vs. Tuples HTML vs. XML

[Erdweg et al, 2010]

Carnegie Mellon University, School of Computer Science

8

Our Solution Libraries cannot extend the base syntax of the language Instead, syntax is associated with types.

Type-specific syntax can be used to create values of that type. How? By placing a tilde (~) where an expression of that

type is expected, and beginning an indented block after the containing declaration/statement.

Carnegie Mellon University, School of Computer Science

9

Example: Architecture Specification

Carnegie Mellon University, School of Computer Science

10

Example: Queries

Carnegie Mellon University, School of Computer Science

11

Examples: HTML and URLs

Carnegie Mellon University, School of Computer Science

12

(provisional syntax)

Type-Associated Grammars

Carnegie Mellon University, School of Computer Science

13

(provisional syntax)

Composition: Wyvern Productions

Carnegie Mellon University, School of Computer Science

14

(provisional syntax)

Composition: Imported Productions

Carnegie Mellon University, School of Computer Science

15

(provisional syntax)

Composition: Typed Wyvern Expressions

Carnegie Mellon University, School of Computer Science

16

Phase I: Top-Level Parsing The top-level layout-sensitive syntax of Wyvern can be

parsed first without involving the typechecker Useful for tools like documentation generators Wyvern’s grammar can be written down declaratively using a layout-sensitive

formalism [Erdweg et al 2012; Adams 2013]

DSL blocks are left as unparsed “DSL literals” during this phase

Carnegie Mellon University, School of Computer Science

17

Phase II: Typechecking and DSL Parsing When a tilde expression (~) is encountered during

typechecking, its expected type is determined via: Explicit annotations Method signatures Type propagation into where clauses

The subsequent DSL literal is now parsed according to the type-associated grammar. Any internal Wyvern expressions are also parsed (I & II) and typechecked

recursively during this phase.

Carnegie Mellon University, School of Computer Science

18

Benefits Modularity and Safe Composability

DSLs are distributed in libraries, along with types No link-time errors

Identifiability Can easily see when a DSL is being used via ~ and whitespace Can determine which DSL is being used by identifying expected type DSLs always generate a value of the corresponding type

Simplicity Single mechanism that can be described in a few sentences Specify a grammar in a natural manner within the type

Flexibility Whitespace-delimited blocks can contain arbitrary syntax

Carnegie Mellon University, School of Computer Science

19

Ongoing Work

Carnegie Mellon University, School of Computer Science

20

Inline DSL Literals Whitespace-delimited blocks admit arbitrary syntax but…

May be unwieldy for simple DSLs (e.g. URLs, times, dates, etc.) Only allow one DSL block per declaration/statement

Solution: Alternative inline forms for DSL literals (with same type-directed semantics) Collection of common delimiter forms

“DSL literal” `DSL literal` {DSL literal} <DSL literal> [DSL literal] /DSL literal/ …

Carnegie Mellon University, School of Computer Science

21

Inline DSL Literals That is, these three forms could be exactly equivalent,

assuming f takes a single argument of type URL f(~)

http://github.com/wyvernlang/wyvern f(`http://github.com/wyvernlang/wyvern`) f([http://github.com/wyvernlang/wyvern]) f(“http://github.com/wyvernlang/wyvern“)

(String literals are simply a DSL associated with the String type!)

Alternatively, types could restrict the valid forms of identifier to allow the language itself to enforce conventions.

Carnegie Mellon University, School of Computer Science

22

Keyword-Directed Invocation Most language extension mechanisms invoke DSLs using

functions or keywords (e.g. macros), rather than types. The keyword-directed invocation strategy can be

considered a special case of the type-directed strategy. The keyword is simply a function taking one argument. The argument type specifies a grammar that captures one or more

expressions.

Carnegie Mellon University, School of Computer Science

23

Example: Control Flowif : bool -> (unit -> a), (unit -> a) -> a

IfBranches

if(in_france, ~) do_as_the_french_do() else panic()

if(in_france) do_as_the_french_do() else panic()

Carnegie Mellon University, School of Computer Science

24

Interaction with Subtyping With subtyping, multiple subtypes may define a grammar. Possible Approaches:

Use only the declared type of functions Explicit annotation on the tilde Parse against all possible grammars, disambiguate as needed Other mechanisms?

Carnegie Mellon University, School of Computer Science

25

Interaction with Tools Syntax interacts with syntax highlighters + editor features. Still need to figure out how to support type-specific syntax

in these contexts. Borrow ideas from language workbenches?

Carnegie Mellon University, School of Computer Science

26

Related Work

Carnegie Mellon University, School of Computer Science

27

Active Libraries [Veldhuizen, 1998]

Active libraries are not passive collections of routines or objects, as are traditional libraries, but take an active role in generating code.

Carnegie Mellon University, School of Computer Science

28

Active Code Completion [Omar et al, ICSE 2012]

Use types similarly to control the IDE’s code completion system.

Carnegie Mellon University, School of Computer Science

29

Active Code Completion with GRAPHITE

Carnegie Mellon University, School of Computer Science

30

Active Code Completion with GRAPHITE

Carnegie Mellon University, School of Computer Science

31

Active Code Completion with GRAPHITE

Carnegie Mellon University, School of Computer Science

32

Active Code Completion with GRAPHITE

Carnegie Mellon University, School of Computer Science

33

Active Typechecking & Translation[Omar and Aldrich, presented yesterday at DSLDI 2013]

Use types to control typechecking and translation.

Implemented in the Ace programming language.

Carnegie Mellon University, School of Computer Science

34

Benefits of DSLs Specialized syntax improves ease-of-use Specialized typechecking rules improve verifiability Specialized translation strategies improve performance

and interoperability with existing technologies Specialized tool support improves ease-of-use.

Carnegie Mellon University, School of Computer Science

35

Types Organize Languages Types represent an organizing principle for programming

languages. Types are not simply useful for traditional verification, but

also safely-composable language-internal extensibility.

Type-Directed, Whitespace-De-limited Parsing for Embedded DSLs

Cyrus Omar

School of Computer Science Carnegie Mellon University

[GlobalDSL13]

Benjamin Chung Alex PotaninDarya Kurilova Jonathan Aldrich

http://cs.c-mu.edu/~dkurilov/

Carnegie Mellon University, School of Computer Science

37

Examples